+ All Categories

imhof

Date post: 15-Jan-2016
Category:
Upload: diana-lujza
View: 214 times
Download: 0 times
Share this document with a friend
Description:
paper
Popular Tags:
16
THEINTL. JOURNAL OF LISTENING, 24: 19-33, 2010 Gopyright © Taylor & Francis Group, LLG ^ g ISSN: 1090-4018 print/ 1932-586X online g X Taylor & Francis Croup DOI: 10.1080/10904010903466295 Listening to Voices and Judging People Margarete Imhof Institute for Psychology Johannes Gutenberg-University The impact of vocal cues on personality judgments is investigated in an experimen- tal study that used technically manipulated levels of pitch (low and high frequency), sex of the speaker, and content area (e.g., fixing a bike, baking, reading directory information) as independent and the entailing personality judgments as dependent variables. Subjects (48 male and 48 female) were presented with voice probes and ratings of physical (age, sex, height, stature), and psychological characteristics (bipolar adjectives representing the "Big Five" dimensions of personality) were col- lected. Results confirm that voice characteristics have an impact on interpersonal perception and that vocal cues are processed separately by the listener. Results are discussed with reference to processing demands and cognitive load on the working memory of a listener. The impact of vocal cues on the listening process would not be challenged by any- one in the field of communication research. Prominent definitions of listening con- tain the aspect that listening includes the perception and interpretation of nonverbal messages (e.g., Purdy, 1997; Wolvin & Coakley, 1996), and nonverbal messages certainly include paralinguistic or "vocal cues that accompany spoken words" (Knapp & Hall, 2002, p, 379). Empirical research has shown that the vocal cues of a message have an infiuence on both how physical and psychological aspects of the person behind the voice are perceived and on how the message is interpreted. More or less subde vocal cues, for example, are believed to disclose both a speaker's temporary aspects, for example, his or her current emotional state (Ellgring & Scherer, 1996; Gobi & Chasaide, 2003; Tischer, 1993) and dispositional character- istics (Brown & Bradshaw, 1985; Krauss, Freyberg, & Morsella, 2002; Scherer, Scherer, Hall, & Rosenthal, 1982), even if reliability and validity of the inferences Correspondence concerning this article should be addressed to Prof. Dr. Margarete Imhof, Johannes Gutenberg-University, Institute for Psychology, Binger Strasse 14-16, D-55099 Mainz, Germany. E-mail: [email protected]
Transcript
Page 1: imhof

THEINTL. JOURNAL OF LISTENING, 24: 19-33, 2010Gopyright © Taylor & Francis Group, LLG ^ gISSN: 1090-4018 print/ 1932-586X online g X Taylor & Francis CroupDOI: 10.1080/10904010903466295

Listening to Voices and Judging People

Margarete ImhofInstitute for Psychology

Johannes Gutenberg-University

The impact of vocal cues on personality judgments is investigated in an experimen-tal study that used technically manipulated levels of pitch (low and high frequency),sex of the speaker, and content area (e.g., fixing a bike, baking, reading directoryinformation) as independent and the entailing personality judgments as dependentvariables. Subjects (48 male and 48 female) were presented with voice probes andratings of physical (age, sex, height, stature), and psychological characteristics(bipolar adjectives representing the "Big Five" dimensions of personality) were col-lected. Results confirm that voice characteristics have an impact on interpersonalperception and that vocal cues are processed separately by the listener. Results arediscussed with reference to processing demands and cognitive load on the workingmemory of a listener.

The impact of vocal cues on the listening process would not be challenged by any-one in the field of communication research. Prominent definitions of listening con-tain the aspect that listening includes the perception and interpretation of nonverbalmessages (e.g., Purdy, 1997; Wolvin & Coakley, 1996), and nonverbal messagescertainly include paralinguistic or "vocal cues that accompany spoken words"(Knapp & Hall, 2002, p, 379). Empirical research has shown that the vocal cues ofa message have an infiuence on both how physical and psychological aspects of theperson behind the voice are perceived and on how the message is interpreted. Moreor less subde vocal cues, for example, are believed to disclose both a speaker'stemporary aspects, for example, his or her current emotional state (Ellgring &Scherer, 1996; Gobi & Chasaide, 2003; Tischer, 1993) and dispositional character-istics (Brown & Bradshaw, 1985; Krauss, Freyberg, & Morsella, 2002; Scherer,Scherer, Hall, & Rosenthal, 1982), even if reliability and validity of the inferences

Correspondence concerning this article should be addressed to Prof. Dr. Margarete Imhof,Johannes Gutenberg-University, Institute for Psychology, Binger Strasse 14-16, D-55099 Mainz,Germany. E-mail: [email protected]

i7
Highlight
i7
Highlight
i7
Highlight
i7
Highlight
i7
Highlight
Page 2: imhof

20 IMHOF

remain problematic (Scherer, 2003). To a certain degree, vocal cues provide hintsas to whether a message is trustworthy (Anolli & Ciceri, 1997; Ekman, Friesen, &Scherer, 1976; Zuckennan, Amidon, Bishop, & Pomerantz, 1982). Voice charac-teristics have also been related to concepts of attractiveness (Collins & Missing,2003; Zuckennan & Miyake, 1993); consequently, much effort goes into designingand training voices for success (Gutzeit, 2002).

Researching vocal cues and their implications for how a person is perceived isa complicated field (Bente & Krämer, 2003; Scherer, 2003). There are basicallytwo avenues that have been taken to describe and explain the relationshipbetween vocal cues and personality:

• Investigate the accuracy with which listeners can identify the "true" charac-teristics of a person whose voice they have heard, for example, how reliablycan one tell the age, sex, education, ethnic group, and a series of psycholog-ical traits (e.g., dominance, extraversion) and states (anxiety, nervousness,trying to tell a lie) from a person's voice. This first line of research has beensuccessful for some aspects of personality perceptions. For example, sex,age, and social class or status can be assessed fairly accurately from a person'svoice after rather brief exposure to the target voice (Krauss, Freyberg, &Morsella, 2002; Sebastian & Ryan, 1985). Also, extraversion and domi-nance as personality dimensions can be inferred from the voice rather con-sistently with test measures (Siegman, 1987). On the whole, however, theobserver's assessment from paralinguistic cues does not correlate tooclosely with the actually measured personality type (Knapp & Hall, 2002).

• Describe the stereotypes that people associate with vocal cues, for example,how are certain voice characteristics (e.g., nasality, loudness) stereotypicallyinterpreted? There is substantial agreement on how vocal cues are interpretedand on what observers believe to be indicators for temporary or dispositionalpersonal characteristics. Studies along these lines reveal, for example, that theinterpretation of the emotional state of a speaker very much depends on thepitch variation and the talking speed (Tischer, 1993). It also has been shownthat attractive voices induce a much more favorable image of the personbehind the voice (Zuckerman & Driver, 1989; Zuckerman & Miyake, 1993).Studies that have used a variety of vocal cues showed that these characteris-tics lead to quite distinct personality perceptions on the side of the listener.There is also a strong differential effect, since identical vocal cues are inter-preted rather differently in a male and female voice (Addington, 1968; seeKnapp & Hall, 2002, p. 389 for an overview). So, for example, a female witha breathy voice is thought to be more feminine, prettier, petite, effervescent,high-strung, and shallower than other females. The same voice characteristicsin a male, however, would lead to a perception of this person as being rela-tively young and more artistic.

i7
Highlight
i7
Highlight
i7
Highlight
i7
Highlight
i7
Highlight
Page 3: imhof

LISTENING TO VOICES AND JUDGING PEOPLE 21

Most Studies, however, are limited in their scope due to one of the followingdesign problems:

• The voice probes are technically manipulated and well-controlled but aredecontextualized and limited in content scope, for example, short individualsentences or even words and syllables entirely without context,

• The voice probes are very close to natural speech, but usually more thanone parameter has been altered; for example, speakers use personalizedwording and differ in more than one of the vocal cues, such as talkingspeed, intonation, micro-pauses, and others, so that the ceteris paribusrequirements are not fully met.

• There is no experimental variation to collect the data for the voice probes.

In the current study, one specific aspect of the voice, that is, pitch, was iso-lated and manipulated as a vocal parameter while keeping all others constant andto control for content and sex of the speaker so that the effects of pitch on the per-ception of the person and the personality of a voice should become visible. Theobjective of this study is to test if the impact of an isolated vocal parameter oninterpersonal perception can be determined and how this effect, if it exists, can bedescribed. The research questions were stated as follows:

• How much does the pitch level of a voice influence the listener's perceptionof the person and the personality behind the voice?

• To what extent do the speaker's sex and the content of an utterance contributeto the listener's perception of the person and the personality behind the voice?

METHOD

Material

After a set of trial runs for a variety of texts and topics, three texts were con-structed for the preparation of the voice probes. These were technically producedfrom two original recordings taken from a male and a female speaker in a sound-proof room. Two of these texts were made to mildly appeal to gender stereo-types, that is, "how to repair the inner tube of a bike" (male) and "how to preparea shortcake" (female). The third text, reading fictitious names and addresses froma directory, was supposed to be neutral in terms of gender specificity. Thus, itwas taken care that both speakers actually used the same wording so that theprobes contain the identical content. The speakers were asked not to read thetexts but practiced to use natural speech as much as possible. Using a voicetransformation program (Wave Lab v4,0), both of these probes were manipulated

Page 4: imhof

22 IMHOF

TABLE 1Pitch Values for the Two Original Voice Recordings and for the Technically Manipulated

Voices in the Voice Probes for the Three Different Texts

How to fix a bike tubeHow to bake a shortcakeReading addresses

Original

218Hz226 Hz2l8Hz

Eemale

High

232 Hz235 Hz233 Hz

Low

197 Hz200 Hz198 Hz

Original

148 Hz154 Hz143 Hz

Male

High

160 Hz165 Hz158 Hz

Low

138 Hz140 Hz137 Hz

into a high and a low pitch version (see Table 1 for the actual frequencies thatwere getierated), the general objective being to maintain all the other speaker andvoice characteristics, for example, speed, modulation and emphasis, pausing,accent, and pronunciation. All voice probes had an average duration of 20-30seconds. This procedure resulted in 12 stimulus variations: Three texts, each real-ized by a male and a female speaker, were transformed into high and low pitchlevel. The actual frequencies of the probes could not be exactly equalizedbecause it had to be taken into account that the resulting speech qualities, forexample, the formants, should not be distorted too much, lest the resulting probemight sound like a synthetic voice. Tlie resulting probes were checked forauthenticity on an informal basis by independent persons, not including theexperimenter and her adviser.

An assessment sheet was used to collect data on the specific impressions thatsubjects had formed of the person behind the voice. Subjects indicated the per-ceived gender (male/female) of the speaker, the age range (< 17; 18-20; 21-23, etc.,in two-year intervals through 35; > 35), height category (< 158 cm; 159-162 cm;163-166 cm, etc., in 3-cm units through > 194 cm), body type (athenic, pyknic,leptosome), speed (on a six-point scale from very slow to very fast), and personalattractiveness (would you like to meet this person? - yes/no). In addition,subjects were asked to use a list of 17 bipolar adjectives to rate the perceivedpersonality type of the speaker. These adjectives were selected to represent thefive major dimensions of personality, known as the "Big Five," that is, extraver-sion, openness, agreeableness, conscientiousness, and emotional stability(Pervin, 2001). A longer list of 37 pairs of antonyms had been tested in a trial runand the items with the most variance and with the least number of missing datawere selected for the final list.

Sample

A total of 96 university students in Germany, 48 males and 48 females, participatedin the study. The average age was 24 with a range from 19 to 38. The subjects were

Page 5: imhof

LISTENING TO VOIGES AND JUDGING PEOPLE 23

all volunteers. They were asked by a student helper if they were willing to partici-pate in an experiment on the assessment of voice. In individual sessions subjectslistened to three out of 12 voice probes that were selected in a way that each subjectwas exposed to all three texts and to two different pitch levels.

Procedure

Subjects were asked individually to participate in the study. They listened to thevoices using headphones from a CD walkman player (SHARP MD-MT866).Each stimulus was presented twice and with the identical preset volume. Beforelistening to the recordings, the subjects had a chance to look over the assessmentsheet which they filled out immediately after the presentation of a voice probehad been completed. The subjects also were instructed on how to use the bipolarscale. Subjects were instructed to focus on the voice and to neglect the contentand other aspects in their judgment. In order to control for serial effects, thesequence of presentation for each subject was determined by random numbers.

RESULTS

The data were processed using SPSS for WINDOWS 10.0. The procedurewas set up to allow for 3 (texts) x 2 (pitch) x 2 (sex) ANOVAs to investigategroup comparisons. The Chi^-test was used for comparisons of categoricaljudgments. Finally, the responses for the 17 pairs of bipolar adjectives werefactor analyzed to obtain a more structured picture of the impressions thatwere reported.

Pitch and Estimates of Speaker's Age

The 3 x 2 x 2 ANOVA procedure using estimated age as the dependent vari-able and content (3), pitch (2), and speaker's sex (2) as fixed factors yieldedboth significant main effects for content and sex as well as significant inter-actions for content and sex (see Table 2). Higher voices were generally attrib-uted to significantly younger speakers. Speakers with higher voices wereassumed to be somewhere between 21 and 23 years old, whereas speakerswith relatively low voices were perceived to be between 27 and 29. Also,female voices were associated with a significantly younger person than malevoices. As shown in Table 2, the effect size was rather large for pitch,whereas it was more moderate for sex. It also turned out that female voicestalking about a male topic (fixing a bike tube) were attributed to significantlyyounger women than the same female voices talking about baking or readingnames.

Page 6: imhof

24 IMHOF

TABLE 2Analysis of Variance for Estimates of Speaker's Age

Source

within subjectspitchcontentsex of the speakerpitch X contentpitch X sex of the speakercontent x sex of the speakererror

df

121212

277

F

232.602.74

11.15.61.09

4.10(1.15)

P

<.000n.s.

<.OOIn.s.n.s.

<.O5

.456-

.039--

.029

Pitch and Estimates of Speaker's Height and Body Type

Table 3 represents the analysis of variance for the estitiiates of the speaker'sheight depending on pitch, the perceived sex, and the content of what he or shesays. Subjects thought speakers with lower voices to be significantly taller thanspeakers with higher voices. The same holds for sex, since the average estimateof the height of a speaker behind a female voice is lower than that of a malevoice. It is notable, however, that the content of what a person talks about has animpact on how tall he or she is represented, since a person describing how to fix abike tube is imagined to be significantly smaller than a person talking aboutbaking or reading names.

As to the physical appearance, subjects envisaged speakers with higher voicesto be more of the athenic body type (slender, long-limbed), whereas speakerswith low voices are thought to be of the pyknic type (shorter, squat, and rounded)(x (N = 288, df = 2) - 51.23, p < .001). Neither content nor perceived sex of the

TABLE 3Analysis of Variance for Estimates of Speaker's Height

Source

within subjectspitchcontentsex of the speakerpitch X contentpitch X sex of the speakercontent x sex of the speakererror

df

121212

276

F

69.826.51

291.01.37.05

1.02(1.52)

P

<.OOI<.O1<.OOI

n.s.n.s.n.s.

.202

.045

.513_--

Page 7: imhof

LISTENING TO VOICES AND JUDGING PEOPLE 2 5

speaker had a significant impact on the listener's representation of the speaker'sstature.

Pitch and Perception of Talking Speed

When judging the talking speed, subjects perceived the higher voices to talk sig-nificantly faster than lower voices (see Table 4). Neither content nor speaker'ssex was used in this ANOVA, because talking speed might not have been con-stant across contents and speakers.

Pitch and Social Curiosity

Subjects responded to the perceived sex and pitch level when they were asked toindicate if they wanted to meet the person behind the voice. They preferredfemale voices over male voices and higher voices over lower ones. The contentabout which they have heard a person talk, however, did not make a difference intheir decision whether they were curious to get to know the person or not. A lin-ear regression analysis was computed to investigate the relative weights of differ-ent predictors for the subjects' decision on whether or not they would like toactually meet the person behind the voice (see Table 5). Results suggest thatpitch and sex of the speaker play an equally important role, whereas content doesnot seem to matter much.

TABLE 4Analysis of Variance tor Estimates of Talking Speed

Source df F p

<.OO1 .042

within subjects

pitcherror

1272

11.80(90)

TABLE 5Linear Regression including Pitch. Speai<er's Sex and Content on tfie Decision of Whether

or not a Listener Wouid Lii<e to Meet the Person Behind a Voice

Variable

pitchsex of speakercontent

Regression Coefficient B

.90-1.13

-.05

Standard Error

.25

.25

.15

Wald

12.72219.854

.095

df

111

P

<.000<.000

n.s.

Exp(B)

2.463.324.954

Page 8: imhof

26 IMHOF

Pitch and Perception of the Personality Behind the Voice

Subjects had rated the personality behind the voice along 17 bipolar adjectives(see Table 6). Ratings were transformed into a five-point scale, so that the neutralpoint (neither the one nor the other) is represented by the value "3"; values below3 indicate a trend toward the end of the bipolar scale, which is represented in thefirst word of the adjective pair. Values above 3 show a trend toward that end ofthe bipolar scale, which is defmed by the second element of the adjective pair.

The multivariate analysis of variance reveals significant main effects for pitchlevel, sex of the speaker, and content (see Table 7), Inspection of between-subjecteffects in individual items shows that the personality ratings are significantlyaffected by pitch level on nearly all listed aspects, except for good-natured vs.irritable, disciplined vs. indulgent, modest vs. ambitious, and open vs. closed.Similarly, most ratings vary with the speaker's sex; exceptions are good-natured VS, irritable, light-headed vs. responsible, not so well educated vs. welleducated, disciplined vs. indulgent, and fragile vs. athletic. The differences inratings, which can be attributed to the content, are not quite as pervading, asthey are significantly visible in only five out of 17 items: mature vs. immature,sluggish vs. dynamic, communicative vs. reserved, disciplined vs. indulgent,and open vs. closed.

To obtain a more structured picture of the results, a principal component fac-tor analysis with varimax rotation was performed. The number of factors to beextracted was determined by the Eigenvalues > 1, This process yielded four fac-tors which together explained 57% of the data variance (see Table 8). Not sur-prisingly, these factors could be interpreted with reference to those fivedimensions of personality that were referred to as the "Big Five" earlier in thisarticle. The factors were named as follows:

• extraversion• openness• conscientiousness• emotional stability• agreeableness

In the present factor analysis, the aspects of extraversion and openness seem tocoincide, which might be due to the specific design of the study, since subjectsonly listened to the voices and had none of the other clues for extraversion thatthey would have in a direct interaction.

Group comparisons based on a full factorial model revealed main effects forpitch, sex of the speaker, and content on one or more dimensions of how thespeaker's personality was assessed. One significant interaction was foundbetween pitch and sex of the speaker (see Table 9),

Page 9: imhof

LISTENING TO VOICES AND JUDGING PEOPLE 2 7

TABLE 6Average Ratings of Speaker's Personality Behind Male and Female Voices

of Different Frequencies Taiking about Different Content Areas

How tofi.x a bike tubeemotional vs. unemotionalgood-natured vs. irritablemature vs. immatureassertive vs. submissivesluggish vs. dynamiclight-headed vs. responsibleattractive vs. unattractivestern vs. lenientnot so well educated vs. well educatedlacking self-confidence vs. self-confidentcommunicative vs. reservedmasculine vs. femininedisciplined vs. indulgentmodest vs. ambitiousfragile vs. athleticopen vs. closedunaffected vs. affected

How to bake a shortcakeemotional vs. unemotionalgood-natured vs. irritablemature vs. immatureassertive vs. submissivesluggish vs. dynamiclight-headed vs. responsibleattractive vs. unattractivestem vs. lenientnot so well educated vs. well educatedlacking self-confidence vs. self-confidentcommunicative vs. reservedmasculine vs. femininedisciplined vs. indulgentmodest vs. ambitiousfragile vs. athleticopen vs. closedunaffected vs. affected

High

M

3.292.252.583.253.503.673.133.423.673.132.833.172.753.252.712.672.21

3.252.293.333.333.333.132.883.833.673.172.172.832.753.382.632.422.38

Male

Voice

SD

1.23.79

1.061.031.18.92.95

1.14.87

1.121.051.051.071.07.86.92.88

1.07.95.76.82

1.05.99.90.70.64

1.051.01.87.99.97

1.101.10.82

Low

M

3..')42.173.543.212.463.833.253.713.543.292.462.042.752.833.752.462.04

3.721.883.723.042.403.723.363.763.683.442.882.402.642.963.442.441.92

Voice

SD

.83

.76

.831.02.88.96

1.07.91.72.91.98.62.99.92.61.88.81

1.06.73.94.98

1.12.89.91

1.13.80

1.041.20.91

1.22.98.92

1.121.04

High

M

2.502.253.212.794.213.332.253.543.583.781.924.422.633.502.751.752.46

2.422.333.333.084.133.382.063.503.383.631.834.582.833.422.712.042.75

Eemale

Voice

SD

.88

.94

.88

.98

.72

.96

.79

.98

.781.001.14.78.11.72.79.90.88

1.06.82.92

1.06.74

1.06.73

1.021.01.88

1.01.58.96.72.81

1.08l.ll

Low

M

3.042.383.792.543.714.042.793.003.834.002.173.922.463.483.042.212.42

3.172.384.002.753.174.002.793.173.633.672.543.292.793.583.542.252.25

Voice

SD

1.08.82.98

1.02.95.69.78.98.64.78.96.88

1.14.95.95.93.93

1.05.82.72

1.071.05.66.98.92.71.96

1.061.00.93.72.72.94.79

(Continued)

Page 10: imhof

28 IMHOF

TABLE 6(Continued)

Reading names and addressesemotional vs. unemotionalgood-natured vs. irritablemature vs. immatureassertive vs. submissivesluggish vs. dynamiclight-headed vs. responsibleattractive vs. unattractivestern vs. lenientnot so well educated vs. well educatedself-conscious vs. self-confidentcommunicative vs. reservedmasculine vs. femininedisciplined vs. indulgentmodest vs. ambitiousfragile vs. athleticopen vs. closedunaffected vs. affected

High

M

3.582.503.003.333.213.923.083.633.502.753.252.962.223.422.543.001.92

Male

Voice

SD

1.10.98.88

1.09.93.78.83

1.06.78

1.15l.ll1.001.00.83.98

1.02.93

Low

M

3.792.173.542.962.384.043.333.334.083.503.292.002.083.463.383.041.75

Voice

SD

.83

.76

.98

.95

.82

.86

.921.01.72.93.86.83.78.88.97.81.85

High

M

2.331.753.333.253.633.381.883.583.793.212.424.462.423.292.382.082.54

Female

Voice

SD

.76

.741.051.07.71.97.90

1.02.93

1.021.10.83.97.95.88

1.06.83

Low

M

2.922.294.042.423.254.132.752.673.963.962.923.792.003.753.382.712.17

Voice

SD

1.32.95.75.93

1.07.68.79

1.09.81.91

1.10.72.83.85.65.91

1.17

TABLE 7Multivariate Analysis of Variance for Effects of Pitch, Speaker's Sex and Content

on Personality Ratings

Source Value Hypothesis df Error df

pitchspeaker's sexcontent

Wilks-LambdaWilks-LambdaWilks-Lambda

12.8724.732.37

171734

256256514

.000

.000

.000

.46

.62

.14

Results suggest that a person with a high voice as compared to a person with alower voice is associated with higher agreeableness on the one hand, anddecreased conscientiousness and emotional stability on the other hand. Femalevoices are in general rated to belong to a person who is more extraverted andopen, whereas male voices are more strongly credited with emotional stabilityand agreeableness. Interestingly, the content of what a person talks about has aninfluence on how the personality of this person is perceived. The personality of aperson who reads addresses is assumed to be significantly less extraverted and

Page 11: imhof

LISTENING TO VOICES AND JUDGING PEOPLE 2 9

TABLE 8Rotated Factor Matrix for Dimension of Personality Ratings

open vs. closedcommunicative vs. reservedattractive vs. unattractivesluggish vs. dynamicemotional vs. unemotionalself-conscious vs. self-confidentlight-headed vs. responsibledisciplined vs. indulgentnot so well educated vs. well educatedunaffected vs. affectedimmature vs. maturemodest vs. ambitiousfragile vs. athleticmasculine vs. femininegood-natured vs. irritablestem vs. lenientassertive vs. submissive

/

Extraversionand

Openness

.784*

.752

.689-.625

.590-.549

-.405

.428

//

Conscientiousness

.740-.702

.654-.591

.548

.537

ttt

EmotionalStability

.304

.396

.442

.761-.729

-.392

tv

Agreeableness

.426

.319

.392

.757-.603-.520

*Factor loadings that meet the Fiimtratt-criterion (a^/h^ > .50) are printed in bold.Factor loadings smaller than .30 are not included.

open than a person whose voice talks about baking or fixing a bike. The samecontent, that is, reading out addresses, however, leads a listener to assume signif-icantly greater conscientiousness. Emotional stability is represented in a persontalking about baking rather than in a person reading out names. Pitch is inter-preted differentially in male and female voices. Women with a low voice are per-ceived to be more agreeable than women with a high voice, whereas the reversepattern emerges for men who are perceived less agreeable when speaking with alow voice.

DISCUSSION AND CONCLUSIONS

Since it had long been known that vocal cues convey certain personal characteris-tics and attitudes, it was of interest to disentangle the net of paralinguistic phe-nomena and to ñnd out more about the effects of pitch as a ubiquitous voicecharacteristic on the listener, that is, on the model that a listener constructs of a

Page 12: imhof

30 IMHOF

TABLE 9Muitivariate Analysis of Variance for Pitch, Speal<er's Sex, and Types of Content

Source

PitchExtraversion and OpennessConscientiousnessEmotional StabilityAgreeableness

Sex of the speakerExtraversion and OpennessConscientiousnessEmotional StabilityAgreeableness

ContentExtraversion and OpennessConscientiousnessEmotional StabilityAgreeableness

ErrorExtraversion and OpennessConscientiousnessEmotional StabilityAgreeableness

df

1111

1111

2222

272272272272

E

12.4011.19

116.05.238

94.891.37

35.717.45

5.906.223.25.53

(.71)(.94)(.64)(.94)

P

<.OOI<.OOI<.OO1

n.s.

<.OOIn.s.

<.OO1<.OI

<.O1<.O1<.O5

n.s.

.04

.04

.30-

.26-.01.12

.04

.04

.02-

speaker's personality from his or her voice. In an experiment, subjects were pre-sented with three out of 12 voice probes varying pitch level with sex of thespeaker and content as additional factors. Subjects were asked to describe therepresentation which they had formed of the person behind the voice. Resultsshowed that pitch actually makes a difference in the way a person is perceived.High pitch levels are generally perceived to belong to individuals who are moreextraverted and open but also convey lower degrees of conscientiousness andemotional stability. In this respect, the results of the current study converge withprevious findings suggesting that higher frequency voices are assessed as moreattractive (Collins & Missing, 2003) and as signaling positive affect (Jay, 2003).This effect could be accounted for by the human capacity to reconstruct correlat-ing face motion from speech acoustics (Yehia, Kuratate, & Vatikiotis-Bateson,2002) so that a mental image might be created from a voice almost automatically.

There is a considerable gender effect in the results. In general, the personalitybehind a female voice is supposed to be more extraverted and open but lessagreeable and weaker in emotional stability than a personality behind a malevoice. This differential effect of male and female voices had been suggested inearlier studies (Whipple & McManamon, 2002) and is supported here. Persons

Page 13: imhof

LtSTENING TO VOtCES AND JUDGtNG PEOPLE 31

talking about a neutral theme are rated as less extraverted and open and strongerin emotional stability and conscientiousness. However, the decision on whetheror not somebody wanted to meet the person behind the voice, pitch and biologi-cal sex of the speaker had the largest impact whereas content did not have a pre-dictive value for this decision.

In addition, it became obvious in this study that listeners take into account thecontent when they judge people. Although the instruction asked participants toneglect the content, they could not entirely tune out what was said. Mehrabian'sconclusion (Mehrabian & Ferris, 1967) that verbal information has only limitedimpact on how speakers are perceived is not supported by the findings of thisstudy. Instead, the results support Knapp and Hall (2002, p, 380) who suggestthat the design of Mehrabian's study which looks only at acoustic expression ofone word might not apply to longer text. The results of this study indicate thatverbal information does have an impact on how listeners perceive speakers.

This study confirms the findings that listeners draw information from per-sonal, verbal, and vocal aspects of the communication and integrate the disparateelements into a unitary impression which, to a certain degree, can be assumed tobe judgmental (Gobi & Chasaide, 2003), On the whole, these findings supportthe concept of the listener as a multitasking agent functioning in multiple modal-ities (Yehia et al,, 2002). The recognition of the cognitive load arising from thenecessity to integrate several sources of information simultaneously needs to beaccounted for in models of speech and discourse perception (Patterson, 1999),

The limitation of this study lies clearly in the fact that there was only one ele-ment of the voice characteristics that was manipulated. Of course, pitch neveroccurs as an isolated variable, and therefore the entire situation was somewhatartificial. The point was, however, to test if the generally accepted assumptionthat paralanguage has an impact on how a message and a person is perceived, issusceptive to empirical investigation. In particular, the question was of interestwhether listeners process specific voice characteristics individually or whetherthey form an amalgamated, wholistic impression from the paralinguistic signals.This latter process does not seem to be an option for the explanation of theimpact of vocal cues. The need for further investigation of the effects of vocalcues increases as the immediacy with which these cues are perceived (Sander &Scheich, 2001) and the sustainable influence of voice characteristics on retention(Karayani & Gardiner, 2003; Mayer, Sobko, & Mautone, 2003) are considered.

Further research should definitely include other vocal characteristics so thatan empirically based functional model of voice effects could be constructed.More research is also needed to check for possible compensatory and synergeticeffects of the interaction of different vocal cues (Scherer, 2003). For the practiceof listening, these results specifically confirm that the working memory of a lis-tener has to come to terms with quite disparate processes, since the acoustic sig-nal needs to be analyzed in at least two respects: the stream of sounds for the

Page 14: imhof

32 IMHOF

verbal cues and the voice for the subtext, such as the speaker's personality andattitudes. It might be hypothesized that listeners are differentially effective withthis analysis, depending on their current or dispositional working memory capac-ity. It is an open question if an effective listener is able to selectively allocateattentional capacity to content and/or the vocal cues in order to make the most ofa message. In any case, it might be helpful for discourse and communicationanalysis to be aware of how vocal cues have an influence on the way in which lis-teners judge and misjudge other people.

REFERENCES

Addington, D. W. (1968). The relationship of selected vocal characteristics to personality perception.Speech Monographs, 35, 492-503.

Anolli, L., & Ciceri, R. (1997). The voice of deception: Vocal strategies of naive and able liars. Jour-nal of Nonverbal Behavior, 21, 259-284.

Bente. G., & Kramer, N. C. (2003). Die integrierte Registrierung verbalen und nonverbalen Kommu-nikationsverhaltens [Integrated processing of verbal and nonverbal aspects of communicationbehavior]. In T. Herrmann & J. Grabowski (Eds.), Enzyklopädie der Psychologie: ThemenbereichC Theorie und Forschung. Serie III Sprache. Band I Sprachproduktion (pp. 219-246). Göttingen,Germany: Hogrefe.

Brown, B. L., & Bradshaw, J. M. (1985). Towards a social psychology of voice variations. InH. Giles & R. N. St. Clair (Eds.), Recent advances in language, communication, and social psy-chology (pp. 144-181). London: Lawrence Erlbaum.

Collins, S. A., & Missing, C. (2003). Vocal and visual attractiveness are related in women. AnimalBehavior, 65, 991-]004.

Ekman, P., Friesen. W. V., & Scherer, K. R. (1976). Body movement and voice pitch in deceptiveInteraction. Semiótica, 16, 23-27.

Ellgring, H., & Scherer, K. R. (1996). Vocal indicators of mood change in depression. Journal ofNonverbal Behavior, 20, 83-110.

Gobi, G., & Ghasaide, A. N. (2003). The role of voice quality in communicating emotion, mood andattitude. Speech Communication, 40, 189-212.

Gutzeit. S. F. (2002). Die Stimme wirkungsvoll einsetzen [How to use your voice effectively].Weinheim: Beltz.

Jay. T. B. (2003). The psychology of language. Upper Saddle River. NJ: Prentice Hall.Karayani, I.. & Gardiner, J. M. (2003). Transferring voice effects in recognition memory from

remembering to knowing. Memory & Cognition, 31, 1052-1059.Knapp, M. L., & Hall, J. A. (2002). Nonverbal communication in human interaction. Fort Worth, TX:

Harcourt Brace College Publishers.Krauss, R. M., Freyberg, R., & Morsella, E. (2002). Inferring speakers' physical attributes from their

voices. Journal of Experimental Social Psychology, 38, 618-625.Mayer, R. E., Sobko, K., & Mautone, P. D. (2003). Social cues in multimedia learning: Role of

speaker's voice. Journal of Educational Psychology, 95, 419-425.Mehrabian, A.. & Ferris. S. R. (1967). Inference of attitudes from nonverbal communication in two

channels. Journal of Counseling Psychology, 31, 248-252.Patterson. M. L. (1999). The evolution of a parallel process model of nonverbal communication. In

P. Philippot, R. S. Feldman, & E. J. Coats (Eds.), The social context of nonverbal behavior(pp. 317-.347). Cambridge, UK: Cambridge University Press.

Page 15: imhof

LISTENING TO VOICES AND JUDGING PEOPLE 3 3

Pervin, L. A. (2001). Personality: Theory and research. New York: Wiley.Purdy, M. (1997). What is listening? In M. Purdy & D. Borisoff (Eds.), Listening in everyday life:

A personal and professional approach (pp. 1-20). Lanham, MD; Llniversity Press of America.Sander. K.. & Scheich. H. (2001). Auditory perception of laughing and crying activates human

amygdala regardless of attentional state. Cognitive Brain Research, 12, 181-198.Scherer, K. R. (2003). Vocal communication of emotion: A review of research paradigms. Speech

Communication, 40, 227-256.Scherer, K. R., Scherer, U., Hall, J. A., & Rosenthal, R. (1982). Die attribution von persönlichkeits-

merkmalen aufgrund auditorischer und visueller hinweisreize [Differential attribution of personal-ity characteristics based on auditory and visual perception]. In K. R. Scherer (Bd.), VokaleKommunikation (pp. 228-252). Weinheim, Germany: Beltz.

Sebastian, R. J., & Ryan, E. B. (1985). Speech cues and social evaluation: Markers of ethnicity, socialclass, and age. In H. Giles & R. N. St. Clair (Eds.), Recent advances in language, communication,and social psychology (pp. 112-143). London: Lawrence Eribaum.

Siegman, A. W. (1987). The telltale voice: Nonverbal messages of verbal communication. In A. W.Siegman & S. Feldstein (Eds.), Nonverbal behavior and communication (2nd ed.). Hillsdale, NJ:Erlbaum.

Tischer, B. (1993). Die vokale Kommunikation von Gefühlen [The vocal communication of emo-tions]. Weinheim, Germany: Psychologie Verlags Union.

Whipple, T. W., & McManamon, M. K. (2002). Implications of using male and female voices incommercials: An exploratory study. Journal of Advertising, 31, 79-91.

Wolvin, A. D., & Coakley, C. G. (1996). Li.uening. Madison, Wl: Brown & Benchmark.Yehia, H. C, Kuratate, T., & Vatikiotis-Bateson, E. (2002). Linking facial animation, head motion

and speech acoustics. Journal of Phonetics, 30, 555-568.Zuckerman, M., Amidon, M. D., Bishop, S. E., & Pomerantz, S. D. (1982). Face and tone of voice in

the communication of deception. Journal of Personality and Social Psychology, 43, 347-357.Zuckerman, M., & Driver, R. E. (1989). What sounds beautiful is good: The vocal attractiveness ste-

reotype. Journal of Nonverbal Behavior, 13, 67-82.Zuckerman, M., & Miyake, K. (1993). The attractive voice: What makes it so? Journal of Nonverbal

Behavior, 17, 119-135.

Page 16: imhof

Copyright of International Journal of Listening is the property of International Listening Association and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.


Recommended